Hospital mortality outcomes are a focus of quality improvement programs conducted by federal and state government agencies and other organizations. Information about hospital mortality outcomes is being collected and reported as part of efforts to improve the quality of health care, by informing the public with reports that compare outcomes across hospitals, and by providing direct financial incentives to hospitals with comparatively better mortality outcomes. Risk adjustment methods are key components of hospital mortality comparisons. Fair comparison of mortality rates across hospitals requires effective adjustments for differences among patients in their baseline mortality risk. Hospital administrative data will soon be supplemented by information distinguishing diagnoses that are present at the time of admission from those that are complications or adverse events that occur during the hospital stay. Information about which diagnoses are present at the time of admission may allow substantial improvements in the validity of hospital mortality comparisons, by eliminating diagnoses representing complications of care from the risk- adjustment algorithms. Preliminary research demonstrates that this information provides a substantial advantage for mortality risk adjustment in studies using administrative data. This proposed research will develop mortality risk adjustment models that make optimal use of present on admission data to adjust for baseline differences among patients. Statistical models using diagnoses reported as present at the time of admission will be developed for 15 hospital mortality measures. The statistical performance achieved by the models will be measured, validated, and compared to competing risk adjustment models. Alternative methods of using the risk adjusted mortality measures to identify hospitals with higher than expected mortality (or lower) will also be evaluated. Finally, two general mortality indices will be developed, which combine information into summary measures.